Re: [patch] bug in cpuid & msr on nosmp machine

[Paul Jackson <pj@sgi.com>]

Jeff wrote:
> but on nosmp machine if the cpu!=0 this procude a BUG();

The line you refer to here is, I believe, in include/linux/cpumask.h:

  #ifdef CONFIG_SMP
    ...
  #else
    ...
  #define cpu_online(cpu)    ({ BUG_ON((cpu) != 0); 1; })
    ...
  #endif

Any thoughts on whether that BUG() is serving any useful purpose?

I'm of a mind to have the non-SMP case of cpu_online(cpu) simply
respond true (1) hardcoded:

  #define cpu_online(cpu)    (1)

Or at least remove the BUG and have it respond true iff cpu == 0:

  #define cpu_online(cpu)    ((cpu) == 0)

While theoretically correct, that BUG() is, so far as I know, of
no redeeming social value.  And it wastes a few bytes of kernel
text space each time cpu_online() is used on an SMP, and causes
the occassional confusion, as in this case.

-- 
                          I won't rest till it's the best ...
                          Programmer, Linux Scalability
                          Paul Jackson <pj@sgi.com> 1.650.933.1373